This research examines children's early knowledge about letters and spelling from a cross-linguistic perspective, looking at children from the U.S., Britain, Australia, Brazil, and Israel. Because learning involves an interplay between the to-be-learned material and the knowledge and predispositions of the learner, one specific aim is to learn more about the patterns in the input for these languages. Using corpora from children's books, we examine aspects of writing that are likely to be salient for young learners, including the lengths of words, the frequencies of letters, and the extent to which spoken words contain phonemes and phoneme sequences that match letter names. The consistency with which phonemes are represented in spelling is also studied. The statistical studies are linked to experimental work on early spelling and letter knowledge in learners of several dialects of English, as well as Portuguese and Hebrew. The behavioral work is motivated by the idea that humans are statistical learners: Knowledge reflects the patterns to which learners have been exposed. Also important are whether a pattern fits with the learner's existing knowledge and the complexity of the pattern. The behavioral work is further motivated by the idea that children bring to bear skills and abilities that are involved in other domains to the learning of literacy. An additional goal of the research is to characterize individual differences in spelling-related knowledge and to determine how early differences relate to later performance. Computerized scoring systems are developed to assess the extent to which young children's spellings reflect the phonological structure of spoken words and the extent to which their spellings honor various nonphonological patterns. These scoring systems are tested in longitudinal studies that examine how characteristics of children's early spellings predict later spelling and reading performance. High levels of literacy are essential in modern technological societies. The results should be useful in the early identification of children who are at risk of problems in learning to spell and read individual words-problems that are often associated with problems in comprehension and written composition. The results should promote the development of spelling tests that consider not only the correctness of the responses, as in typical standardized tests, but also the nature of the errors. The findings will help educators and researchers take advantage of the information that children's spellings provide.